Tag: infrastructure

Categories
AI Business

The Wage of Knowing

In 1973 the Los Angeles Public Library installed a telephone line that worked while the building was dark. Dial H-O-O-T-O-W-L on a rotary phone, nine at night until one in the morning, and a librarian would answer. Somebody wanted to know the boiling point of mercury, or who wrote a poem they half remembered, or how many wives Henry VIII actually had, and a person on the other end of a cord found out. This went on for years. Nobody thought of it as data collection. It was just a service, a courtesy, a woman at a desk with a card catalog in her head.

I worked, in another life, in the payments industry, back when a merchant who wanted to charge your card had to call in and ask permission. There were rooms for this. Banks of phones, a bulletin of stolen numbers updated by hand, a floor limit past which a supervisor had to be found. The people answering the phones were, more often than you would guess, college students. Twenty years old, minimum wage, deciding in real time whether a stranger’s card was good. Nobody trained them for six months first. They learned the bulletin, they learned to listen for something wrong in a voice, and they said yes or no.

I have been driven, recently, by a car with nobody driving it. I noticed the wheel turning on its own and I braced for the wrongness of it. Thirty seconds later I was not bracing. I was looking out the window. The data says I was right to relax: across two hundred and twenty million miles, the cars involved in this experiment cause a small fraction of the serious crashes a human would have caused over the same roads. I did not need the data. I needed thirty seconds.

None of these people knew what they were doing. That is the thing about the librarian and the college student and, for that matter, about me learning to trust a wheel that moves by itself. The librarian was not building a search engine. The clerk was not training a fraud model. He was making rent. Their competence was not evidence, to them. It was just Tuesday. It became evidence later, to someone else, in a room they never saw — the accident logs, the chargeback data, the accumulated record of a million correct guesses that turned out to be exactly the material a system needed to learn the job and take it.

This is the part that is easy to get wrong. It is not that the human failed and the machine succeeded. It is that the human succeeding, over and over, in full view, was the demonstration that the job could be learned. You do not automate a task nobody can do. You automate the one being done well enough, often enough, for long enough that the pattern becomes visible. Doing the job right was never neutral. It was the case being built.

Which brings me to a woman I will call the lawyer, because there are thousands of her and none of them are exactly her. She has a laptop open at her kitchen table. She logs into a dashboard belonging to a company that pairs credentialed people with the AI labs that need them — a doctor here, a banker there, a corporate attorney with fifteen years of contract law behind her. She reads a model’s draft of a merger agreement and marks where it reasons like a first-year associate instead of a partner. She rewrites a clause. She explains, in the margin, why the model’s version would get laughed out of a negotiation. She is paid well for this. More, some weeks, than she billed certain clients.

She knows exactly what she is doing. That is the difference between her and the other three. The librarian did not know she was leaving a trail. The clerk did not know his good judgment would become someone else’s weights. I did not know, thirty seconds into that ride, that I was participating in anything at all. The lawyer knows. She is being paid, by the hour, at a rate that respects her expertise, to make her expertise legible enough that it no longer requires her. The company she works for has a name for this. They call it the reinforcement learning economy, which is a tidy way of saying: teach it everything, and then it will not need to call you back.

She does the work anyway. The rate is good. The work is interesting, in the way that teaching is interesting — you learn what you know by trying to say it clearly enough for someone else to use. Nobody is lying to her. The dashboard does not pretend to be anything other than what it is. She logs off at the end of the session the way anyone logs off after a long day of being excellent at something, tired in the specific way that comes from careful work, and she does not, from what I understand, spend the evening thinking about what she has just fed into the machine.

I keep coming back to the rotary dial. Somebody dialing H-O-O-T-O-W-L at midnight in 1973 could not have imagined the lawyer at her kitchen table. But the shape is the same, if you look at it long enough. A person answers a question well. The answering becomes a record. The record becomes a system. The system answers next time. Nobody in the room ever decided this was the plan. It just turned out, every time, to be the plan.

Categories
AI Apple Google

The Floor

I compared the frontier to a three-star chef making grilled cheese in “Context Rot” — the smartest models on earth spending most of their time on work beneath them, the way a chef trained at Le Bernardin might still melt cheese between two slices of bread on a Tuesday night and call it dinner. The comfort was the point: if the sharpest tool is saved for hard problems and something merely-very-good handles the rest, nobody’s losing anything. The floor was never the interesting part.

I’ve kept turning the joke over, and I think I had the wrong worry.

Watch what companies do with their AI spend, not what they say. Coinbase moved engineers off frontier models onto open weights and cut its AI spend nearly in half while usage kept climbing. Nvidia runs a closed model as orchestrator and routes the actual volume — the daily uncelebrated bulk of it — to open weights it controls. The frontier is becoming a dispatcher, deciding where the request goes and rarely doing the work itself. The instinct is to worry about whose open weights end up running that volume, and right now the most capable ones at scale are Chinese — GLM, Kimi — which makes it tempting to read this as a contest America is quietly losing: the floor of the AI economy built somewhere else, at a price export controls can’t touch. You cannot embargo a file already downloaded. You cannot price-match free.

But that framing has a hole. Google’s own Gemma family is open-weight and good enough to handle that daily volume without anyone reaching for GLM or Kimi. “Open weights are a Chinese story” only holds if you don’t count the open models the company running Android and half the internet’s search traffic has already shipped.

And once I saw that hole, a bigger one opened behind it. I’ve been trying Apple’s new Siri — arriving with iOS 27 this fall, genuinely surprisingly good in beta — and it made me realize open weights, of any nationality, were never going to cook most of the world’s dinners. Apple and Google are.

Consider what actually determines where the world’s routine inference runs. Not which model benchmarks best, not which weights are downloadable — what’s already installed. Apple ships to well over a billion active devices before routing a single query through Siri’s new architecture. Nobody has to be persuaded to try it, or hear about it on a podcast; it’s the thing that answers when you press the button you’ve pressed for a decade. Google owns the search bar and the Android default the same way. Between them, that’s most of the world’s phones — and phones are where most of the world’s questions get asked.

The open-weight framing assumes the floor is up for grabs, that whoever ships the best free model wins the daily grind by merit. But the floor was never a bazaar. It’s a set of defaults, owned by whoever already has the device in your hand, not whoever holds the most generous license. Apple didn’t need to win the model war to win this. Its heaviest reasoning tier is built with Google, running on Nvidia chips in Google’s cloud, under a deal reported at roughly a billion dollars a year — Apple doesn’t fully own the engine doing the thinking. It doesn’t need to. It owns the button.

That’s a quieter concentration than an export-controls fight, and a harder one to dislodge. An open model can be forked, distilled, undercut, or out-competed by the next release. A billion phones with an assistant built into the lock screen cannot be routed around. Whoever’s weights hum underneath barely matters, the way it barely matters to a diner which supplier delivered the flour. What matters is whose kitchen the meal came from, and whose name is on the door.

The grilled-cheese chef was never the risk. Two chefs are about to own nearly every kitchen on earth, and most of us will never notice — because a kitchen you’ve been eating out of for a decade doesn’t feel like something that was won. It just feels like home.

Owning the kitchen and getting paid for what’s cooked in it, though, turn out to be two different questions. That one’s for another post.

Categories
Design Technology

The Battery That Refused to Leave

A standard AA battery is 50.5 millimeters long and 14.5 millimeters in diameter. It produces 1.5 volts. It weighs roughly twenty-three grams, about as much as a sheet of paper folded twice. In a Costco bulk pack, forty-eight of them together weigh a little over a kilogram — the heft of a hardcover book, or a decent cantaloupe. Most people buy them without thinking much about it. They go in the cart the way paper towels go in the cart.

The size has been in continuous production since 1907, when the American Ever Ready Company first manufactured it for use in early penlights. For the first four decades of its existence, the AA battery was what might be called an informal standard — widely used, commonly understood, but not officially codified. That changed in 1947, when the American National Standards Institute fixed the dimensions and voltage in writing. The naming convention itself had come earlier, out of a series of meetings in the 1920s between government officials and battery manufacturers who were trying to bring order to a proliferating market. They began with A for the smallest practical cell, then moved outward — B, C, D — for larger sizes. When smaller cells were needed later, the alphabet doubled back on itself: AA, AAA, AAAA. Running out of letters in both directions is its own kind of history.

What the standards committee built, whether they thought of it this way or not, was a commons. The word is precise. A commons is something no one owns and everyone can use — a pasture, a fishery, a language. The AA battery became a commons of power. Any battery from any manufacturer, made to the specification, would work in any device built to receive it. The chemistry inside could vary — zinc-carbon, alkaline, lithium, nickel-metal hydride — but the housing stayed the same. No license was required. No negotiation. A manufacturer building a flashlight in 1965 did not need to solve the battery problem. A company making a remote control in 1985 did not need to negotiate with a power supplier. The relationship between a device and its energy source belonged to no one, which meant it was available to everyone.

In 1959, an Eveready scientist developed the first commercially available alkaline AA, which lasted five to eight times longer than the zinc-carbon version it was designed to replace. The devices followed the power. Transistor radios. Portable tape players. Handheld games. Cameras. Each decade brought a new category of device that found the AA battery waiting for it, already standardized, already available at every drugstore and grocery checkout lane in the country. The commons kept growing because the commons was free to enter.

Apple, eventually, decided the idea was wrong.

The iPhone, introduced in 2007, had no user-replaceable battery. Neither did any iPod before it, any iPad after it, any MacBook, any AirPod, any Apple Watch. The power source in an Apple product is sealed inside the device, charged through Apple’s own cables and connectors, managed by Apple’s own software. This is not a cost-cutting measure or an engineering compromise. Apple’s products cost more than their competitors’, not less, and the sealed battery is part of what justifies the price. The company’s founding argument — refined over decades, made explicit in every product announcement — is that hardware and software and power, designed together and optimized together, produce a better result than any open standard can achieve. The AA battery asks nothing of you except that you insert it correctly. Apple has decided that is insufficient.

Tesla arrived at a similar conclusion by a different route. Where Apple sealed the power source to improve the user experience, Tesla sealed it to own the energy relationship entirely. The Supercharger network — Tesla’s proprietary charging infrastructure, built out across highways and cities at enormous expense — is not interoperable with other electric vehicles, or was not for most of its history. A Tesla charges at a Tesla station. The battery chemistry, the cell format, the thermal management, the software that governs charging and discharge — all of it is developed in-house, at Tesla’s gigafactories, for Tesla’s vehicles. The company has spent more time and money thinking about batteries than almost any organization outside of a national laboratory. But the battery it produces is not a commodity. It belongs to the car. The car belongs to Tesla’s ecosystem. The customer belongs there too.

Both companies are making a version of the same argument: that the future of technology is integrated, that the best products are closed products, that power should be managed rather than swapped. They have built that future, or a version of it, for the customers who can afford to live inside it.

Warren Buffett, in 2014, bought the thing neither of them wanted.

Berkshire Hathaway’s acquisition of Duracell from Procter & Gamble was structured as a stock swap — Berkshire exchanged its $4.7 billion stake in P&G for full ownership of the battery company, recapitalized with $1.8 billion in cash. The tax advantages were real and significant; Berkshire had held the P&G shares since the company’s acquisition of Gillette in 2005, and the cost basis was $336 million. A cash sale would have produced a substantial capital gains bill. The swap avoided that. Buffett is attentive to such things.

But the more durable rationale was simpler. Buffett has spent sixty years looking for businesses that are easy to understand, that generate predictable cash, that sell something people buy out of habit. See’s Candy. GEICO. Coca-Cola. The common thread is not glamour but persistence — products whose value proposition does not need to be reinvented, whose customers return not because they have been excited but because they have been satisfied, reliably, for a long time. Duracell has twenty-five percent of the global battery market. It has been the category leader for decades. The people who buy it at Costco are not making a considered choice between competing technologies. They are buying what they have always bought.

The Costco pack of forty-eight is, in Buffett’s framework, infrastructure. Not the infrastructure of data centers or power grids — the quiet infrastructure of daily life, the kind that gets restocked when the supply runs low and otherwise goes unnoticed. Smoke detectors. Remote controls. Children’s toys. Wireless computer mice. Clocks on kitchen walls. The devices that run on AA batteries are not going away, and the economics of replacing them — not just the devices but the habits, the muscle memory, the universal availability of the standard — are formidable. Buffett is not betting that the AA battery will conquer the future. He is betting that it will remain in the present for a very long time.

Two different visions of where technology is going, then, expressed in the form of capital allocation. Apple and Tesla have built sealed ecosystems and asked their customers to enter. Buffett bought the battery for the people who haven’t. The AA cell, fifty millimeters long and fourteen and a half millimeters wide, 1.5 volts, unchanged in its dimensions since a group of manufacturers met in the 1920s to agree on something everyone could use — it sits at the back of a kitchen drawer in most houses in America, waiting for the smoke detector to chirp.

Categories
Writing

The Grain Bin and the Ghost

Richard Rhodes published How to Write in 1995. In it, he offers practical advice about a writer’s reference shelf: keep a dictionary at home, own a one-volume encyclopedia. He mentions, almost in passing, that he just received the OED on CD-ROM as a birthday gift.

That sentence stops you cold in 2026.

Not because it’s quaint — though it is — but because of what it reveals about how a writing life was organized. Rhodes wasn’t describing luxury. He was describing infrastructure. The reference shelf was load-bearing. You kept facts at home the way you kept food in a pantry: because access wasn’t guaranteed, because the library closed, because the gap between not-knowing and knowing could be measured in trips and hours. A writer’s bookshelf was a personal hedge against scarcity.

Think about what it meant that someone’s birthday present was a reference tool. Not a novel. Not a bottle of wine. Twenty volumes of the most authoritative dictionary in the English language, compressed to a disc, given because a writer needed it and couldn’t otherwise have it on their desk. That’s what a writing life cost. That’s what it demanded of the people around you.

That scarcity is gone so completely it’s hard to reconstruct the phenomenology of it.

The bottleneck in Rhodes’s world was access. You either had the fact or you didn’t. Getting it required physical movement — to the shelf, to the library, to someone who knew. The reference book’s value was proximity: it collapsed the distance between the question and the answer. The OED on CD-ROM was remarkable precisely because it put those twenty volumes on your desk. No trip. No waiting. That was the gift.

The bottleneck now is entirely different. Access is solved, trivially, for anyone with a phone. The question isn’t where the facts are. The question is which facts to trust, how they were assembled, whether the source has an agenda, whether the model that synthesized them has introduced drift. We moved from a scarcity problem to a judgment problem, and most of our inherited intellectual habits were built for the former.

But something else happened too, something Rhodes couldn’t have framed because it didn’t exist: the infrastructure became generative. The reference shelf held facts. It didn’t think with you. It didn’t draft alongside you, or push back on your argument, or notice that the claim you just made contradicts something three paragraphs earlier. The CD-ROM OED was static; it waited to be consulted. The tools a writer reaches for now don’t wait. They participate.

This is the structural shift that the grain bin metaphor can’t contain. Rhodes was describing a writer’s relationship to stored knowledge — how you accumulate it, how you keep it close, how you move through it when you need it. That relationship was essentially curatorial. You built a collection. You maintained it. You drew from it.

What’s emerging now is something more like a collaboration with an infrastructure that has opinions. Not always right ones. Not always trustworthy ones. But opinions nonetheless — which means the writer’s job has changed in kind, not just in degree. You’re no longer managing a pantry. You’re managing a working relationship.

Where does it end up? Probably somewhere Rhodes would recognize at the level of the goal — clarity, the right word, the true sentence — and find almost unrecognizable at the level of method. The shelf is still there. But it talks back now. And figuring out what that means — whether to trust it, when to push against it, how to stay the one doing the writing — is the work no one has finished yet. Maybe no one can, while it’s still changing this fast.

Categories
Aging AI Business Living

The Being Phase

There is a metric making the rounds in technology investing circles that is, on its face, about market share and revenue concentration. Alex Sacerdote of Whale Rock Capital calls it the New Rule of 40 for AI. The formula is simple: take the percentage of a company’s sales derived from AI, add its percentage market share in that AI category, and if the sum reaches 40, you have a winner. Celestica, a company most people have never heard of, scores extraordinarily well. It owns somewhere between half and sixty percent of the cloud Ethernet white-box switch market. NVIDIA doesn’t need a formula. It simply is what it is.

Sacerdote designed the metric to cut through a specific kind of noise — the companies claiming AI exposure they don’t actually have, the giants whose AI revenue hovers at one or two percent of their base while their press releases suggest otherwise. The framework is a detector. It finds the companies that have stopped becoming AI infrastructure and started simply being it.

I found myself less interested in the companies than in that distinction.

I spent years at Visa watching a network that had long since crossed that threshold. By the time I arrived, Visa wasn’t becoming the global payments infrastructure. It was the global payments infrastructure. The work was real — fraud detection, modeling, the daily labor of keeping something enormous running — but the existential question had been settled before I got there. The network existed. Merchants accepted it because cardholders carried it. Cardholders carried it because merchants accepted it. That loop had been closing for decades. We were custodians of a fait accompli.

There’s a particular feeling to working inside something that has already won. It’s not complacency exactly. The problems are genuine and the stakes are high. But the uncertainty has a different quality — it’s operational uncertainty, not existential uncertainty. You’re not asking whether the thing will survive. You’re asking how to run it well.

I didn’t have language for that distinction then. Sacerdote’s metric gives me some. The companies that score highest on his New Rule of 40 have resolved their existential question. They’re not fighting for position. They’re administering a position already held.

The question that has followed me out of that career, and out of several decades of watching technology cycles turn, is simpler and more personal than any investment framework.

When did I cross that line myself?

I have been writing at sjl.us since 2001. That’s not a boast — it’s a data point. Twenty-five years of thinking out loud, of ideas arriving rather than being argued, of the specific memory as structural anchor. The blog is not becoming anything. It is what it is: a record of a mind moving through time, accumulated into something that has its own weight and shape.

The book on payments systems exists. The career at Visa exists. The photographs exist. The train journeys exist. The years in Dayton exist, and the years on the Peninsula, and the particular way the light falls on the California coast at Pescadero in the late afternoon — when the fog is still offshore and the hills are improbably green and everything goes briefly, completely quiet, as if the world is deciding whether to continue.

These are not things I am building toward. They are things I am.

Sacerdote would say I have high market share in a specific category. The category is small — one person, one particular configuration of experience and attention and accumulated knowing — but the share is essentially total. There is no competitor for the position of having lived this particular life. The moat is absolute. The switching costs are infinite.

I used to find that thought melancholy. The narrowing as loss. The aperture closing on what remains.

I’m not sure I find it melancholy anymore.

The L-Curve, Sacerdote says, is a long flatline followed by a vertical explosion. The tinkering phase, then the moment of lift. He means it as a description of demand curves for technology infrastructure. But I recognize the shape from somewhere closer. The long middle of a life, building and becoming, and then the morning you wake up and realize the building is substantially done. What remains is the being.

That’s not an ending. It’s a different kind of beginning.

Sacerdote’s metric will eventually stop working. All frameworks do. The AI infrastructure cycle will mature, the L-Curves will flatten, and some new measure will emerge to find the next thing that is just beginning to become what it will be. That’s the nature of markets. The detector has to change as the signal changes.

But there’s a complication worth naming. Analysts at Citadel Securities published a note recently observing that even the most powerful technologies must pass through the prosaic discipline of cost curves, capacity constraints, and marginal returns. Token bills are arriving unexpectedly. Compute is scarce. The vision of AI as ubiquitous, frictionless, and immediate is colliding with physical reality. Their conclusion: asset prices will periodically be forced to reconcile ambition with physical constraint.

That’s not a refutation of Sacerdote. It’s a reminder that feeling like you’ve arrived and having actually arrived are different things. The being phase has to be load-tested. The position has to hold under pressure.

I think about the fiber optics Corning is laying into the massive data center clusters — ultra-thin, bendable, carrying more light than anything that came before. The cable doesn’t know it’s infrastructure. It just carries what it’s given, at the speed it’s capable of, across whatever distance is required. It doesn’t matter what the cable believes about itself. What matters is whether the light actually moves.

That seems right to me. You become what you are over a long time, largely without noticing. And then one day someone builds a metric that accidentally describes your life, and you recognize yourself in it, and you think: yes. That’s the shape of it. High concentration. High share. A moat that deepened while you were looking elsewhere.

But the moat still has to hold.

The being phase, it turns out, is not the end of something. It’s the proof that something was built. And the daily question — for companies, for infrastructure, for a person in his late seventies still writing, still paying attention — is whether what was built is actually load-bearing.

You don’t get to stop finding out.

Categories
Technology

The Silence of Glass

There is a moment, right before surgery, when the anesthesiologist asks you to count backward from ten. You get to seven, maybe six, and then the world goes clean and white. Scientists have a word for the material responsible for that transition: borosilicate. The same compound in the syringe barrel is in the telescope mirror trained on the Andromeda galaxy, in the fiber strand carrying the surgeon’s consultation with a colleague three thousand miles away, in the smartphone screen the patient’s wife is staring at in the waiting room, hands shaking, refreshing nothing.

Glass is everywhere and we have made it invisible, which is the oldest trick civilization knows.

Vaclav Smil argues in Making the Modern World that the most consequential material of the last two centuries is not steel or silicon or oil. It is float glass — invented by Alastair Pilkington in 1959, when he watched dishwater spread across his kitchen sink and understood something that had eluded glassmakers for four hundred years. Pour molten glass onto a bath of molten tin and it finds its own level. It becomes, on its own, perfectly flat. Every window, phone screen, solar panel, and architectural facade descends from a man watching his wife do dishes.

What Smil doesn’t quite say — though you feel it accumulating across his pages — is that glass is the one material that consistently mediates between the inner and the outer. Not metaphorically. Literally. It stands at the boundary and says: you may look, but you may not touch.

The fiber optic cable looks like nothing. Pull back the orange jacket and you find strands thinner than a human hair, each one pure silica glass so precisely drawn that a photon launched into one end will emerge after sixty miles having lost less than five percent of its energy. That number seems impossible. It is a kind of miracle achieved through obsessive purity: any contaminant at the molecular level, any stress in the crystal lattice, any deviation in the core diameter, and the light scatters and dies. Underneath every ocean, through every mountain, connecting data centers in Virginia to servers in Singapore, there are hundreds of millions of kilometers of this material, laid in darkness, carrying light.

I think about that sometimes when I hit send. The electrons leave my keyboard, convert to photons at some local junction, and then travel — genuinely travel, as light through glass — to wherever they are going. There is something devotional about it, though I can’t quite say why. Maybe it’s the invisibility. Maybe it’s the faith required — that the thing you release will arrive, intact, somewhere it has never been.

Glass is in the MRI machine and the X-ray plate and the laboratory flask where the drug was first synthesized and the vial where it is stored and the syringe through which it enters the body. Glass does not react. It does not corrode. It does not leach. This chemical inertness, which seems like absence, is actually the whole point. Medicine needed a container that would hold the thing without becoming it.

There is also glass in the eye reading the label on that vial. The human lens is, optically speaking, a soft glass. It focuses, ages, clouds — cataracts are the eye’s glass going milky — and the surgeon replaces it with an intraocular lens engineered to behave like glass. We have spent considerable effort making fake versions of something the body was already doing.

For most of human history, clear glass was expensive, fragile, and small. Window glass in medieval Europe admitted light hazily, like looking through ice. Clear vision was for churches, which is perhaps why we came to associate light with the sacred — it literally arrived, in those buildings, in a way it did not arrive anywhere else. Then Pilkington’s tin bath made clarity cheap, and the world changed in ways nobody fully catalogued because the change was so pervasive: big windows, watched experiments, extended growing seasons, telescopes reaching farther, microscopes going smaller. Each a story of glass making a distance crossable that was not crossable before.

The screen I am writing this on is glass. The Corning Gorilla Glass on this display is an alkali-aluminosilicate sheet, chemically strengthened through ion exchange, harder than most knives, clear enough that the pixels look like they are sitting on the surface rather than behind it. Apple spends considerable engineering effort making the glass seem like it isn’t there. The ideal phone screen is invisible. A window to computation.

And yet the glass is the thing you actually touch. All day. More than you touch almost anyone. The glass is warm from your hands. It has learned, in a way, the pressure of your thumbs.

Glass is the material of thresholds — it makes the threshold visible, makes it possible to stand at a door and see all the way through before you decide whether to enter. We built the internet through it. We see our loved ones through it. We study cancer through it. We watch the news through glass that traveled to us through glass captured by cameras with glass sensors launched on satellites with glass lenses through a sky that is itself, technically, a lens — bending and filtering the light from everything that has ever been.

In the hospital waiting room, the wife is still holding her phone. The screen has gone dark. She taps it. It lights up. She looks at her own reflection for a moment — the screen a mirror now — before the notification arrives and the glass goes transparent again, the way it always does, showing her something other than herself.

That is what glass does. It waits. It holds. And then, when there is something to show, it gets out of the way.

Categories
AI AI: Large Language Models China

Cranes on the Horizon

In 2005, during my first trip to Shanghai and Beijing, the most striking feature of the skyline wasn’t the architecture—it was the cranes. More than I could possibly count, perched atop half-finished skyscrapers like a mechanical forest. Entire districts seemed to be mid-construction simultaneously, as if someone had pressed a button and the whole country decided to build everything at once. Dan Wang in his book “Breakneck” described China as the “engineering state” that approaches national problems with physical solutions. Back in 2005, coming from Silicon Valley, I thought I understood what growth looked like. I didn’t.

I’ve been thinking about that trip while reading Nathan Lambert’s recent piece, “Notes from Inside China’s AI Labs.” Lambert — who runs the Interconnects newsletter and does serious work tracking the open-weight LLM ecosystem — just returned from visiting essentially every major AI lab in China. Moonshot, Zhipu, Meituan, Xiaomi, Qwen, Ant Ling, 01.ai. He went in with genuine curiosity and came back with humility. That combination is rarer than it should be.

What he found was the cranes. Different domain, same energy.

Lambert’s central observation is about culture, not capability. The Chinese labs aren’t winning on any single technical breakthrough — they’re winning on execution discipline. He describes researchers, many of them active students, who bring no ego to the work. They absorb context fast, drop assumptions faster, and seem genuinely unbothered by the philosophical debates that seem to swirl constantly in the American AI community. When he tried to engage Chinese researchers on the long-term social risks of models or the ethics of AI behavior, those questions “hung in the air with a simple confusion. It’s a category error to them.” Their role is to build the best model. Full stop. To them, an LLM isn’t a philosophical entity to be interrogated; it’s a piece of infrastructure to be optimized.

That description landed for me. Not as a criticism of American research culture, but as a real observation about what the moment demands. Building good LLMs today is, as Lambert puts it, meticulous work across the entire stack — “all points of the model can give some improvements, and fitting them in together is a complex process.”

The work that matters most right now isn’t the 0-to-1 creative leap; it’s the thousand unglamorous decisions executed without complaint. Students who haven’t yet learned to lobby for their own ideas turn out to be well-suited for exactly this.

Lambert ends on a note that’s hard to shake. Looking up from his laptop on a high-speed train, he keeps seeing cranes on the horizon. He draws the same connection I did, though from the inside: the construction everywhere fits the broader culture and energy around building. “When I look up from my laptop and always see bunches of cranes on the horizon, it obviously fits in with the broader culture and energy around building in China.”

Twenty years after my first visit, the cranes are still there. They’ve just moved indoors — into server rooms and training runs and model releases that land every few months with quiet confidence. In 2005, what China was building was obvious: you could see the steel frames going up. What’s being built now is harder to see, which may be exactly why it keeps surprising us.

Check out Lambert’s essay – it’s remarkable. If the 20th century was defined by who could move the most earth, the 21st will be defined by who can move the most tokens. And right now, the cranes are moving faster than we think.

Categories
California Petroleum

The Last Tanker

There is a strange, quiet finality to the arrival of the New Corolla. It is a massive vessel, carrying two million barrels of crude—a literal, physical weight of energy—into the Port of Long Beach. It loaded up in Iraq on February 24th, just days before the world’s geopolitical plates shifted and the Strait of Hormuz effectively slammed shut.

By the time you read this, that oil will have been offloaded, refined, and moved into the capillaries of California’s infrastructure—into gas tanks, jet engines, and diesel generators.

And then, the silence begins.

California has long existed as an “energy island.” It is a geographic quirk that defines our modern life: we are disconnected from the domestic pipeline network that feeds the rest of the country. We don’t have the luxury of pulling from a pipeline in Texas or the Midwest. We are, by design, tethered to the horizon. We are dependent on the flow of tankers across the vast, deep blue of the Pacific.

For years, this worked. It was a invisible architecture of convenience. We consumed, and the tankers arrived with the metronomic precision of a clock. But the New Corolla is not just a delivery; it is a period at the end of a sentence. It represents the last of a supply chain that we assumed would be permanent.

When the analyst says, “all bets are off,” they aren’t just talking about prices at the pump or the logistical scrambling of refineries trying to source crude from Brazil or Guyana. They are describing the erosion of a certainty we didn’t realize we relied on. We have built a state—a massive, humming, technological engine—on the assumption that the world is a frictionless marketplace.

The crisis is not just about the supply of oil; it is the realization that we are fragile.

We look at our inventories, and we see them as a buffer. We are told they are “healthy,” but inventories are, by definition, a countdown. They are the water left in the glass after the tap has been turned off. We are now in the uncomfortable, interim phase where the supply lines are empty, and the new ones haven’t yet been built—or perhaps, cannot be built.

It is easy to look at this and see a political or economic failure. It is harder to see it as a human one. We have become experts at consuming the distant, while remaining strangers to the mechanics of that consumption. We have lived in the architecture of the “global everything,” and now, as the walls of that architecture contract, we are forced to look at the geometry of our own isolation.

The New Corolla will depart for distant waters. It will leave behind a void, and in that void, we will find out if our resilience is as robust as our rhetoric.

The future is only guaranteed for those who can afford to survive the present.

And for now, the present is a question of how much gasoline is left in the tank, how much jet fuel is available and how quickly we can learn to walk on our own.

Categories
Interstate 280 San Francisco/California San Jose

The Scenic Route Home

“In a world optimized for speed and engagement, 280 is a reminder that infrastructure can be art.”

It is a strange paradox that in the heart of Silicon Valley—a place defined by the ephemeral, the digital, and the instantaneous—a cherished shared experience is a physical ribbon of highway that hasn’t changed much in fifty years.

My post from last April, “The World’s Most Beautiful Freeway,” has recently found a new wave of readers. I’ve been asking myself: Why? Why does a blog post about Interstate 280, written by a retiree exploring local history, resonate so deeply right now?

Perhaps it’s because I-280 is more than just a commute. As I noted in the original piece, even Sunset Magazine in 1967 recognized it as “a modern and scenic boulevard.” It was a bold claim for a freeway, yet it stuck. While its sibling, US 101, is a clogged artery of billboard-choked utility, 280 feels like a deep breath. It is the “scenic route” we are lucky enough to take right in our own backyard.

There is a powerful nostalgia in that drive. We all remember the sign that used to sit in the median near Cupertino—the one that literally proclaimed it “The World’s Most Beautiful Freeway”—before it vanished. We remember the way the fog rolls over the Santa Cruz Mountains, spilling into the crystal bowl of the reservoir.

But I think the recent interest goes deeper than pretty scenery. We are living in an era of rapid, often disorienting change. I used ChatGPT to help research the history of that road, a small testament to how AI is weaving into our daily inquiries. Yet, the road itself remains a constant. It was designed by engineers like Othmar Ammann and planners who chose the harder, more expensive route through the foothills rather than paving over El Camino Real. They chose beauty over pure efficiency.

That choice resonates today. In a world optimized for speed and engagement, 280 is a reminder that infrastructure can be art. It connects the headquarters of the companies building our future (Apple, Google, Meta) with the wild, golden hills of California’s past. It is a physical timeline of the Peninsula.

Maybe we are revisiting this post because we are craving that balance. We want to know that even as we rush toward the future at freeway speeds, we can still look out the window and see something timeless, something beautiful, something that reminds us where we are.

Categories
AI Software

The Thermodynamics of Thought

For the last two decades, we have lived in the era of zero marginal cost. The defining characteristic of the internet age was that once software was written, distributing it to the billionth user cost virtually the same as distributing it to the first. We grew accustomed to the economics of abundance—infinite copies, infinite reach, lightweight infrastructure.

But the recent commentary regarding the true nature of Artificial Intelligence forces a jarring mental correction:

“AI is not software riding on old infrastructure. It is a new industrial system that converts energy into intelligence – requiring a capital stack measured in trillions, not billions.”

This distinction is not merely semantic; it is physical.

When we view AI through the lens of traditional SaaS (Software as a Service), we miss the magnitude of the shift. We are looking for an app; what is being built is a refinery. We are witnessing a return to heavy industry, but the commodity being refined isn’t crude oil—it is information, and the byproduct is reasoning.

This requires us to think less in terms of code and more in terms of thermodynamics. In this new industrial system, intelligence is an energy-intensive output. Every token generated, every inference drawn, requires a specific, measurable conversion of electricity into heat and computation. Unlike the static code of a website, an AI model is a furnace. It must be fueled constantly.

This explains the capital stack. We are seeing numbers that seem irrational in the context of venture capital—trillions, not billions. But if you view a data center not as a server farm, but as a power plant that generates intelligence, the numbers align with historical precedents. We are not funding startups; we are funding the modern equivalent of the electric grid, the transcontinental railroad, or the petrochemical complex.

We are pouring concrete, smelting copper, and manufacturing silicon on a planetary scale. The “cloud” was always a misleading metaphor—it sounded fluffy and ethereal. The reality of the AI transition is heavy, hot, and incredibly expensive.

We are moving from an era where we organized the world’s information (low energy) to an era where we synthesize new reasoning (high energy). We are building a machine that eats electricity and excretes intelligence. That isn’t a software update; that is a new industrial revolution.